News for hackers. Stuff that matters.
Re:ChatGPT: The Slickest Con Artist of All Time by b3morales on Friday February 03, 2023 @11:50AM (#34643707)
> help me practice a foreign language I've slowly been forgetting

So with the software stuff you can pretty easily verify the output: either the suggestion fixes the problem or it doesn't. But how can you trust it for something like this where you can't distinguish what's good and bad output? It could be leading you down the garden path with this language exercise. And it's not in the helpful thinking tool/bouncing ideas around category, either: there are rules of grammar that you don't know, and you don't know whether ChatGPT knows them either.

There are no comments on this comment yet.